(032) Exploring the Use of Electronic Health Records (EHR) Data for Active Safety Surveillance of Epidiolex in PCORnet, July 1, 2018- July 1, 2022

Publication Number: 1028

Background: In June 2018, the US Food and Drug Administration (FDA) approved the first cannabis-derived cannabidiol (CBD) human drug product (Epidiolex), currently indicated for the treatment of seizures associated with Lennox-Gastaut syndrome, Dravet syndrome, or tuberous sclerosis complex in patients ages ≥1 year.

Objectives: To explore the use of PCORnet for active safety surveillance of Epidiolex.

Methods: We conducted this descriptive analysis in the FDA’s Sentinel System using data from 23 DataMarts in PCORnet®, an electronic health records-based network of networks that uses a common data model. We required patients to have ≥1 encounter prior to a CBD exposure (day 0) between July 1, 2018, through July 1, 2022, to define the following exposure-based cohorts: (1) Epidiolex and (2) CBD (no brand name). We assessed demographics, clinical characteristics (days -365, 0), medications (days -30, 30), and healthcare utilization (days -365, -1).

Results: The Epidiolex cohort included 6,306 individuals with a mean age of 27 (SD: ±17) years, 51% female, and mostly White (82%). Of them, 31% had a diagnosis for which Epidiolex is indicated and 66% for other epilepsies. The CBD cohort comprised 2,558 individuals with a mean age of 37 (SD:±24) years, 55% females, 74% White; 20% had a diagnosis for which Epidiolex is indicated and 50% other epilepsies. Overall, clinical characteristics were similar between cohorts, with slightly lower prevalence of cardiovascular disease (13% vs 16%), cancer (5% vs 9%), and hypertension (12% vs 15%), and slightly higher prevalence of behavioral/emotional disorders (18% vs 13%), anorexia/weight loss (6% vs 4%), and sleep disorders (20% vs 17%) in the Epidiolex cohort than in the CBD cohort. We observed similar proportion of records for anticoagulants, antidiabetics, CNS stimulants, chemotherapies, systemic corticosteroids, other immunosuppressants, and opioids whereas the Epidiolex cohort had higher record of anticonvulsants (71% vs 50%), antihistamines (18% vs 10%), and anxiolytics (30% vs 16%). As per healthcare utilization, 57% and 40% had ≥6 records in the Epidiolex and CBD cohorts, respectively.

Conclusions: Overall, both cohorts did not differ by demographics, clinical characteristics, medications, or healthcare utilization suggesting that most CBD exposures identified actually correspond to Epidiolex. Some of the differences might be explained by the younger mean age in the Epidiolex cohort. Although further explorations are required, results suggest that, when using PCORnet for active surveillance of Epidiolex, we should not overlook CBD exposures without brand name.
Disclaimer: The contents are those of the authors and do not necessarily represent the official views of, nor an endorsement, by FDA/HHS, or the US Government.